SVM Classifiers 18C 16x
Uriah Daugaard
4/29/2021
1 Info
Here I train the 16x video classifiers with the merged data, i.e. using stock culture data from every time point and light setting recorded.
2 Dependencies
library(tidyverse)
library(randomForest)
library(readr)
library(viridis)
library(e1071)
library(caret)
library(patchwork)3 Data: load in and cleaning
Note: there are only few individuals of the small ciliates because only cells bigger than 1000um2 are tracked at this magnification and most small ciliates (Tetrahymena, Dexiostoma and Loxocephalus) are smaller than that.
3.1 Normal light
3.1.1 1st classifier/data (2020-2021)
Note: there are only few individuals of the small ciliates because only cells bigger than 1000um2 are tracked at this magnification and most small ciliates (Tetrahymena, Dexiostoma and Loxocephalus) are smaller than that.
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Coleps_irchel/Morph_mvt.RData")
dd16 <- morph_mvt
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Colpidium/Morph_mvt.RData")
morph_mvt$species <- "Colpidium"
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Euplotes/Morph_mvt.RData")
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Euplotes_second/Morph_mvt.RData")
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Paramecium_bursaria/Morph_mvt.RData")
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Paramecium_caudatum/Morph_mvt.RData")
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Stylonychia_1/Morph_mvt.RData")
morph_mvt$species <- "Stylonychia1"
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Stylonychia_1_second/Morph_mvt.RData")
morph_mvt$species <- "Stylonychia1"
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Stylonychia_2/Morph_mvt.RData")
morph_mvt$species <- "Stylonychia2"
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Dexiostoma/Morph_mvt.RData")
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Loxocephallus/Morph_mvt.RData")
dd16 <- rbind(dd16, morph_mvt)
load("../../Class_18C_normalLight/1st_class_2020/data/16x/Tetrahymena/Morph_mvt.RData")
morph_mvt$species <- "Tetrahymena"
dd16 <- rbind(dd16, morph_mvt)
# filtering
dd16 <- dd16 %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16$magnification <- 1.6
# table(dd16$species)3.1.2 2nd classifier data (2022 Feb) - normal light
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Coleps_irchel/Morph_mvt.RData")
dd16_2022feb <- morph_mvt
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Colpidium/Morph_mvt.RData")
morph_mvt$species <- "Colpidium"
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Dexiostoma/Morph_mvt.RData")
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Euplotes/Morph_mvt.RData")
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Loxochephalus//Morph_mvt.RData")
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Paramecium_bursaria/Morph_mvt.RData")
# morph_mvt %>%
# ggplot(aes(mean_area))+
# geom_histogram()+
# geom_vline(xintercept = 3000)
morph_mvt <- morph_mvt %>% dplyr::filter(mean_area>3000)
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Paramecium_caudatum/Morph_mvt.RData")
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Stylonychia2_batch1/Morph_mvt.RData")
morph_mvt$species <- "Stylonychia2_batch1"
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Stylonychia2_batch2/Morph_mvt.RData")
morph_mvt$species <- "Stylonychia2_batch2"
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
load("../../Class_18C_normalLight/2nd_class_2022Feb/data/16x/Tetrahymena/Morph_mvt.RData")
dd16_2022feb <- rbind(dd16_2022feb, morph_mvt)
# filtering
dd16_2022feb <- dd16_2022feb %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16_2022feb$magnification <- 1.6
# table(dd16_2022feb$species)3.1.3 3rd data (20220301) - normal light
load("../../Class_18C_normalLight/3rd_data_20220301/data/16x/large_ciliates/Morph_mvt.RData")
dd16_20220301 <- morph_mvt
load("../../Class_18C_normalLight/3rd_data_20220301/data/16x/small_ciliates/Morph_mvt.RData")
dd16_20220301 <- rbind(dd16_20220301, morph_mvt)
# filtering
dd16_20220301 <- dd16_20220301 %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16_20220301$magnification <- 1.6
# dd16_20220301 %>%
# ggplot(aes(mean_area))+
# geom_histogram()+
# facet_wrap(~species, scales = "free")
dd16_20220301 <- dd16_20220301 %>%
dplyr::filter(!(species %in% c("Paramecium_caudatum","Paramecium_bursaria","Euplotes") & mean_area<2000),
!(species %in% c("Dexiostoma","Tetrahymena","Loxocephallus") & mean_area>1500))
# table(dd16_20220301$species)3.1.4 4th data (20220316) - normal light
load("../../Class_18C_normalLight/4th_data_20220316/data/16x/large_ciliates/Morph_mvt.RData")
dd16_20220316 <- morph_mvt
load("../../Class_18C_normalLight/4th_data_20220316/data/16x/small_ciliates/Morph_mvt.RData")
dd16_20220316 <- rbind(dd16_20220316, morph_mvt)
# filtering
dd16_20220316 <- dd16_20220316 %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16_20220316$magnification <- 1.6
# dd16_20220316 %>%
# ggplot(aes(mean_area))+
# geom_histogram()+
# facet_wrap(~species, scales = "free")
dd16_20220316 <- dd16_20220316 %>%
dplyr::filter(!(species %in% c("Paramecium_caudatum","Paramecium_bursaria","Euplotes") & mean_area<2000),
!(species %in% c("Dexiostoma","Tetrahymena","Loxocephallus") & mean_area>1400))
# table(dd16_20220316$species)3.1.5 5th data (20220406) - normal light
load("../../Class_18C_normalLight/5th_data_20220406/data/16x/large_ciliates/Morph_mvt.RData")
dd16_20220406 <- morph_mvt
load("../../Class_18C_normalLight/5th_data_20220406/data/16x/small_ciliates/Morph_mvt.RData")
dd16_20220406 <- rbind(dd16_20220406, morph_mvt)
# filtering
dd16_20220406 <- dd16_20220406 %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16_20220406$magnification <- 1.6
# dd16_20220406 %>%
# ggplot(aes(mean_area))+
# geom_histogram()+
# facet_wrap(~species, scales = "free")
dd16_20220406 <- dd16_20220406 %>%
dplyr::filter(!(species %in% c("Paramecium_caudatum","Paramecium_bursaria","Euplotes") & mean_area<2000),
!(species %in% c("Dexiostoma","Tetrahymena","Loxocephallus") & mean_area>1400))
# table(dd16_20220406$species)3.2 Decreasing light
3.2.1 2nd classifier data (2022 Feb) - 18 perc
load("../../Class_18C_decreasingLight/Light_18perc/data/16x/Euplotes/Morph_mvt.RData")
dd16_2022feb_decrease <- morph_mvt
load("../../Class_18C_decreasingLight/Light_18perc/data/16x/Paramecium_bursaria/Morph_mvt.RData")
a <- morph_mvt %>%
ggplot(aes(mean_area))+
geom_histogram()
Pburs_dark_16x <- morph_mvt %>%
dplyr::select(mean_area,sd_area,mean_perimeter,sd_perimeter,mean_major,
sd_major,mean_ar,sd_ar,
mean_turning,sd_turning,gross_disp,max_net,net_disp,net_speed,max_step,
min_step,sd_step,sd_gross_speed,max_gross_speed,min_gross_speed)
set.seed(23)
Pburs_dark_16x_clust <- kmeans(Pburs_dark_16x, centers = 2, nstart = 25, iter.max = 1000)
Pburs_dark_16x$cluster <- as.factor(Pburs_dark_16x_clust$cluster)
b <- Pburs_dark_16x %>%
ggplot(aes(mean_area, col=cluster, fill=cluster))+
geom_histogram(position = "identity", alpha=0.3)
Pburs_dark_16x_clust <- kmeans(Pburs_dark_16x, centers = 3, nstart = 25, iter.max = 1000)
Pburs_dark_16x$cluster <- as.factor(Pburs_dark_16x_clust$cluster)
c <- Pburs_dark_16x %>%
ggplot(aes(mean_area, col=cluster, fill=cluster))+
geom_histogram(position = "identity", alpha=0.3)
Pburs_dark_16x_clust <- kmeans(Pburs_dark_16x, centers = 4, nstart = 25, iter.max = 1000)
Pburs_dark_16x$cluster <- as.factor(Pburs_dark_16x_clust$cluster)
d <- Pburs_dark_16x %>%
ggplot(aes(mean_area, col=cluster, fill=cluster))+
geom_histogram(position = "identity", alpha=0.3)
morph_mvt <- morph_mvt %>%
dplyr::mutate(cluster=as.factor(Pburs_dark_16x_clust$cluster)) %>%
dplyr::filter(cluster !="2") %>%
dplyr::select(-cluster)
e <- morph_mvt %>%
ggplot(aes(mean_area))+
geom_histogram()
# a + b + c + d + e + plot_annotation(title = "Cleaning of P.burs dark 16x")
dd16_2022feb_decrease <- rbind(dd16_2022feb_decrease, morph_mvt)
# filtering
dd16_2022feb_decrease <- dd16_2022feb_decrease %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16_2022feb_decrease$magnification <- 1.6
# table(dd16_2022feb_decrease$species)3.2.2 3rd data (20220301) - 10 perc
load("../../Class_18C_decreasingLight/Light_10perc/data/16x/large_ciliates/Morph_mvt.RData")
dd16_20220301_decrease <- morph_mvt
load("../../Class_18C_decreasingLight/Light_10perc/data/16x/small_ciliates/Morph_mvt.RData")
dd16_20220301_decrease <- rbind(dd16_20220301_decrease, morph_mvt)
# filtering
dd16_20220301_decrease <- dd16_20220301_decrease %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16_20220301_decrease$magnification <- 1.6
# dd16_20220301_decrease %>%
# ggplot(aes(mean_area))+
# geom_histogram()+
# facet_wrap(~species, scales = "free")
dd16_20220301_decrease <- dd16_20220301_decrease %>%
dplyr::filter(!(species %in% c("Paramecium_bursaria","Euplotes","Paramecium_caudatum") & mean_area<2500),
!(species %in% c("Coleps_irchel") & mean_area<1000),
!(species %in% c("Dexiostoma","Tetrahymena","Loxocephallus") & mean_area>1500))
# table(dd16_20220301_decrease$species)3.2.3 4th data (20220316) - 6 perc
load("../../Class_18C_decreasingLight/Light_6perc/data/16x/large_ciliates/Morph_mvt.RData")
dd16_20220316_decrease <- morph_mvt
load("../../Class_18C_decreasingLight/Light_6perc/data/16x/small_ciliates/Morph_mvt.RData")
dd16_20220316_decrease <- rbind(dd16_20220316_decrease, morph_mvt)
# filtering
dd16_20220316_decrease <- dd16_20220316_decrease %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16_20220316_decrease$magnification <- 1.6
# dd16_20220316_decrease %>%
# ggplot(aes(mean_area))+
# geom_histogram()+
# facet_wrap(~species, scales = "free")
dd16_20220316_decrease <- dd16_20220316_decrease %>%
dplyr::filter(!(species %in% c("Paramecium_bursaria","Euplotes","Paramecium_caudatum") & mean_area<2000),
!(species %in% c("Coleps_irchel") & mean_area<1000),
!(species %in% c("Dexiostoma","Tetrahymena","Loxocephallus") & mean_area>1400))
# table(dd16_20220316_decrease$species)3.2.4 5th data (20220406) - 1 perc
load("../../Class_18C_decreasingLight/Light_1perc/data/16x/large_ciliates/Morph_mvt.RData")
dd16_20220406_decrease <- morph_mvt
load("../../Class_18C_decreasingLight/Light_1perc/data/16x/small_ciliates/Morph_mvt.RData")
dd16_20220406_decrease <- rbind(dd16_20220406_decrease, morph_mvt)
# filtering
dd16_20220406_decrease <- dd16_20220406_decrease %>%
filter(net_disp>50, net_speed>5, N_frames>10)
dd16_20220406_decrease$magnification <- 1.6
# dd16_20220406_decrease %>%
# ggplot(aes(mean_area))+
# geom_histogram()+
# facet_wrap(~species, scales = "free")
dd16_20220406_decrease <- dd16_20220406_decrease %>%
dplyr::filter(!(species %in% c("Paramecium_bursaria","Euplotes","Paramecium_caudatum") & mean_area<2000),
!(species %in% c("Coleps_irchel") & mean_area<1000),
!(species %in% c("Dexiostoma","Tetrahymena","Loxocephallus") & mean_area>1400))
# table(dd16_20220406_decrease$species)4 Further data Processing
4.1 Merge data
dd16$species <- ifelse(dd16$species %in% c("Tetrahymena","Dexiostoma","Loxocephallus"),
"Smaller_ciliates",dd16$species)
dd16_2022feb$species <- ifelse(dd16_2022feb$species %in% c("Tetrahymena","Dexiostoma","Loxocephallus"),
"Smaller_ciliates",dd16_2022feb$species)
dd16_20220301$species <- ifelse(dd16_20220301$species %in% c("Tetrahymena","Dexiostoma","Loxocephallus"),
"Smaller_ciliates",dd16_20220301$species)
dd16_20220301_decrease$species <- ifelse(dd16_20220301_decrease$species %in% c("Tetrahymena","Dexiostoma","Loxocephallus"),
"Smaller_ciliates",dd16_20220301_decrease$species)
dd16_20220316$species <- ifelse(dd16_20220316$species %in% c("Tetrahymena","Dexiostoma","Loxocephallus"),
"Smaller_ciliates",dd16_20220316$species)
dd16_20220316_decrease$species <- ifelse(dd16_20220316_decrease$species %in% c("Tetrahymena","Dexiostoma","Loxocephallus"),
"Smaller_ciliates",dd16_20220316_decrease$species)
dd16_20220406$species <- ifelse(dd16_20220406$species %in% c("Tetrahymena","Dexiostoma","Loxocephallus"),
"Smaller_ciliates",dd16_20220406$species)
dd16_20220406_decrease$species <- ifelse(dd16_20220406_decrease$species %in% c("Tetrahymena","Dexiostoma","Loxocephallus"),
"Smaller_ciliates",dd16_20220406_decrease$species)
dd16$data <- "Late 2020"
dd16_2022feb$data <- "20220201"
dd16_2022feb_decrease$data <- "20220201"
dd16_20220301$data <- "20220301"
dd16_20220301_decrease$data <- "20220301"
dd16_20220316$data <- "20220316"
dd16_20220316_decrease$data <- "20220316"
dd16_20220406$data <- "20220406"
dd16_20220406_decrease$data <- "20220406"
dd_30perc <- rbind(dd16,dd16_2022feb,
dd16_20220301 %>% dplyr::select(-Light_dark),
dd16_20220316 %>% dplyr::select(-Light_dark),
dd16_20220406 %>% dplyr::select(-Light_dark))
dd_18perc <- dd16_2022feb_decrease
dd_10perc <- dd16_20220301_decrease %>%
dplyr::select(-Light_dark)
dd_6perc <- dd16_20220316_decrease %>%
dplyr::select(-Light_dark)
dd_1perc <- dd16_20220406_decrease %>%
dplyr::select(-Light_dark)
dd_30perc$light <- "30%"
dd_18perc$light <- "18%"
dd_10perc$light <- "10%"
dd_6perc$light <- "6%"
dd_1perc$light <- "1%"
dd <- rbind(dd_30perc,dd_18perc,dd_10perc,dd_6perc,dd_1perc)
dd$species <- ifelse(dd$species == "Stylonychia_2", "Stylonychia2", dd$species)
dd$species <- ifelse(dd$species %in% c("Stylonychia2_batch1","Stylonychia2_batch2"), "Stylonychia2", dd$species)
# table(dd$species, dd$data, dd$magnification, dd$light)4.2 Filtering out outliers
If an individual is an outlier in more than 3 traits it gets excluded (about 7% gets excluded). Outlier based on boxplot definition.
dd$id <- 1:nrow(dd)
dd <- dd %>% na.omit()
dd_long <- dd %>%
dplyr::select(species, mean_area,sd_area,mean_perimeter,sd_perimeter,mean_major,
sd_major,mean_ar,sd_ar,
mean_turning,sd_turning,gross_disp,max_net,net_disp,net_speed,max_step,
min_step,sd_step,sd_gross_speed,max_gross_speed,min_gross_speed, id,
data, light, magnification) %>%
pivot_longer(cols=-c(id,species,data,light,magnification), names_to="variable") %>%
dplyr::group_by(variable,species,data,light,magnification) %>%
mutate(iqr = IQR(value),
first_quartile = quantile(value, probs = 0.25),
third_quartile = quantile(value, probs = 0.75),
outlier = ifelse(value<first_quartile-1.5*iqr | value>third_quartile+1.5*iqr,T,F))
outliers <- dd_long %>%
dplyr::filter(outlier==T) %>%
dplyr::group_by(species, id, data, light, magnification) %>%
dplyr::summarise(n = n()) %>%
dplyr::filter(n>3)
# table(outliers$species)
# nrow(outliers)/nrow(dd)
dd_filtered <- dd %>%
dplyr::filter(!is.element(id,outliers$id))
dd$removed_outliers <- F
dd_filtered$removed_outliers <- T
dd_comparison <- rbind(dd,dd_filtered)
dd <- dd_filtered
# dd_comparison %>%
# dplyr::filter(magnification==1.6) %>%
# ggplot(aes((mean_area), col=removed_outliers))+
# geom_histogram(aes(fill=removed_outliers, y=..density..), position = "identity", alpha=0.3) +
# # geom_boxplot(outlier.alpha = 0.3) +
# theme_bw() +
# facet_wrap(species~interaction(light,data), scales = "free")
#
# dd_filtered %>%
# ggplot(aes(log10(mean_area)))+
# geom_density(aes(col=species))
# table(dd$species, dd$data, dd$magnification, dd$light)4.3 Sub-sampling to create training data
We have data from several time points and light settings. The goals is that in the final training data these are somewhat equally represented, regardless that the number of individuals tracked per species and time point can vary a lot (in other words: “even it out” across species, time point and light setting). This is important, as otherwise it can be that groups that are more abundant are weighted more.
dd <- dd %>%
mutate(data.light = interaction(data,light, drop = T),
data.light.species=interaction(data,light,species, drop = T))
print("number of individuals per species per date and light setting")## [1] "number of individuals per species per date and light setting"tab <- table(dd$data.light, dd$species)
tab##
## Coleps_irchel Colpidium Euplotes Paramecium_bursaria
## 20220406.1% 43 189 69 332
## 20220301.10% 1421 462 68 140
## 20220201.18% 0 0 1404 1464
## 20220201.30% 2038 559 476 960
## 20220301.30% 1072 402 106 141
## 20220316.30% 207 436 286 1461
## 20220406.30% 133 199 236 763
## Late 2020.30% 1120 296 1879 2307
## 20220316.6% 329 468 697 1375
##
## Paramecium_caudatum Smaller_ciliates Stylonychia1 Stylonychia2
## 20220406.1% 199 2 0 237
## 20220301.10% 744 352 0 922
## 20220201.18% 0 0 0 0
## 20220201.30% 1134 75 0 1149
## 20220301.30% 1308 494 0 2071
## 20220316.30% 1563 21 0 1243
## 20220406.30% 566 19 0 1017
## Late 2020.30% 2038 30 561 347
## 20220316.6% 1371 33 0 925print("Sum of individuals per species")## [1] "Sum of individuals per species"colsums <- colSums(tab)
colsums## Coleps_irchel Colpidium Euplotes Paramecium_bursaria
## 6363 3011 5221 8943
## Paramecium_caudatum Smaller_ciliates Stylonychia1 Stylonychia2
## 8923 1026 561 7911Of course, Stylo1 is the one we the least individuals, because it went extinct and we do not have it anymore. As it went extinct, it shouldn’t be the limiting factor, I think. For now at least I’m taking 2 * Stylo1_sum as the number of individuals per species to be included (if a species/class has less than that all individuals are included).
Within a class I use that number so that roughly equally many individuals across time steps and light settings are included. As a last thing I divide the data into a training dataset and in a test dataset.
n <- min(colsums)*2
print(paste0("Max number of individuals used per species (if there are fewer for a species, then all are used): n=",n))## [1] "Max number of individuals used per species (if there are fewer for a species, then all are used): n=1122"num_ind_per_class <- dd %>% group_by(species) %>%
summarize(num_class = length(unique(data.light.species)),
num_ind_per_class = floor(n/num_class)) %>%
select(species,num_ind_per_class)
num_ind_per_class_vec <- c(num_ind_per_class$num_ind_per_class)
names(num_ind_per_class_vec) <- num_ind_per_class$species
set.seed(53)
split_up <- split(dd, f = dd$data.light.species)
split_up <- lapply(split_up, function(x) {
species <- unique(x$species)
x %>% sample_n(ifelse(nrow(x) < num_ind_per_class_vec[species], nrow(x), num_ind_per_class_vec[species]))})
trainingdata <- do.call("rbind", split_up)
print("Subsampled: number of individuals per species per date and light setting")## [1] "Subsampled: number of individuals per species per date and light setting"tab2 <- table(trainingdata$data.light, trainingdata$species)
tab2##
## Coleps_irchel Colpidium Euplotes Paramecium_bursaria
## 20220406.1% 43 140 69 124
## 20220301.10% 140 140 68 124
## 20220201.18% 0 0 124 124
## 20220201.30% 140 140 124 124
## 20220301.30% 140 140 106 124
## 20220316.30% 140 140 124 124
## 20220406.30% 133 140 124 124
## Late 2020.30% 140 140 124 124
## 20220316.6% 140 140 124 124
##
## Paramecium_caudatum Smaller_ciliates Stylonychia1 Stylonychia2
## 20220406.1% 140 2 0 140
## 20220301.10% 140 140 0 140
## 20220201.18% 0 0 0 0
## 20220201.30% 140 75 0 140
## 20220301.30% 140 140 0 140
## 20220316.30% 140 21 0 140
## 20220406.30% 140 19 0 140
## Late 2020.30% 140 30 561 140
## 20220316.6% 140 33 0 140print("Subsampled: Sum of individuals per species")## [1] "Subsampled: Sum of individuals per species"colsums2 <- colSums(tab2)
colsums2## Coleps_irchel Colpidium Euplotes Paramecium_bursaria
## 1016 1120 987 1116
## Paramecium_caudatum Smaller_ciliates Stylonychia1 Stylonychia2
## 1120 460 561 1120trainingdata <- trainingdata[complete.cases(trainingdata),]
# A stratified random split of the data
idx_train <- createDataPartition(trainingdata$species,
p = 0.75, # percentage of data as training
list = FALSE)
testdata <- trainingdata[-idx_train,]
trainingdata <- trainingdata[idx_train,]5 Loading in species compositions
species <- unique(dd$species)
compositions <- read_csv("../../compositions.csv")## Rows: 15 Columns: 20## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (1): composition
## dbl (19): richness, Chlamydomonas, Cryptomonas, Monoraphidium, Cosmarium, St...##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.comp_name <- compositions$composition
compositions$Smaller_ciliates <- with(compositions,
ifelse(Tetrahymena == 1| Dexiostoma == 1| Loxocephallus == 1, 1, 0))
compositions <- compositions[,species]
compositions_list <- apply(compositions, 1, function(x){
idx <- which(x==1)
names(idx)
})
names(compositions_list) <- comp_name
## filterting out Stylo1
compositions_list <- lapply(compositions_list, function(x){
x[x!="Stylonychia1"]
})6 Classifiers
6.1 Classifier plotting functions
classifier_plot_svm <- function(table, combination.nr){
# cm <- classifier$confusion
cm <- table
ncol <- ncol(cm)
cm <- apply(cm, 1, function(x) x/sum(x))
cm_long <- data.frame(Predicted = factor(rep(rownames(cm),ncol), levels = rownames(cm)),
Truth = factor(rep(colnames(cm), each=ncol), levels = colnames(cm)),
Classification = as.vector(cm))
plot <- cm_long %>%
ggplot(aes(Predicted,Truth,fill=Classification))+
geom_tile() +
geom_text(aes(label = round(Classification, 3)), col="white") +
scale_x_discrete(position = "top") +
scale_y_discrete(limits = rev(levels(cm_long$Truth))) +
scale_fill_viridis(option = "D", end=.8, discrete=FALSE)+
theme(axis.text.x = element_text(angle = 90, hjust = 0))+
theme(legend.text = element_text(size=14),
axis.title = element_text(size=14),
title = element_text(size=16),
axis.text = element_text(size=13))+
labs(title=paste("PPV of",combination.nr),fill="PPV")
return(plot)
}
classifier_assessment_plot <- function(cf, combination.nr){
cf.df <- cf$byClass[,1:4] %>%
as.data.frame() %>%
mutate(Species = gsub("Class: ", "", rownames(cf$byClass[,1:5]))) %>%
rename("NPV" = "Neg Pred Value", "PPV" = "Pos Pred Value",
"Sens" = "Sensitivity", "Spec" = "Specificity") %>%
pivot_longer(cols = 1:4) %>%
mutate(col = ifelse(value>=0.9,"1",
ifelse(value<0.9 & value>=0.8,"2",
ifelse(value<0.8 & value>=0.7,"3","4"))))
plot <- cf.df %>%
ggplot(aes(name,Species,fill=col))+
geom_tile() +
geom_text(aes(label = round(value, 3)), col="white") +
scale_x_discrete(position = "top") +
scale_y_discrete(limits = rev(levels(as.factor(cf.df$Species)))) +
scale_fill_manual(values = c("#006837","#66bd63","#f46d43","#a50026"), breaks = c("1","2","3","4")) +
theme(legend.text = element_text(size=14),
title = element_text(size = 16),
axis.title = element_blank(),
axis.text = element_text(size=13),
legend.position = "none")+
labs(title=combination.nr, fill="")
return(plot)
}6.2 Training of classifiers
formula <- factor(species) ~ mean_area + sd_area + mean_perimeter + sd_perimeter + mean_major +
sd_major + mean_ar + sd_ar +
mean_turning + sd_turning + gross_disp + max_net + net_disp + net_speed + max_step +
min_step + sd_step + sd_gross_speed + max_gross_speed + min_gross_speed
set.seed(7456)
classifiers_18c_16x <- list()
classifiers_18c_16x_plots <- list()
classifiers_18c_16x_assess_plots <- list()
for(i in 1:length(compositions_list)){
sub_trainingdata <- trainingdata %>%
filter(species %in% c(compositions_list[[i]]))
sub_testdata <- testdata %>%
filter(species %in% c(compositions_list[[i]]))
sub_trainingdata$species <- factor(sub_trainingdata$species)
sub_testdata$species <- factor(sub_testdata$species)
# split_up <- split(sub_trainingdata, f = sub_trainingdata$species)
# subsamples <- lapply(split_up, function(x) {
# x %>% sample_n(ifelse(nrow(x) < 500, nrow(x), 500))})
# sub_trainingdata <- do.call("rbind", subsamples)
# A stratified random split of the data
# idx_train <- createDataPartition(sub_trainingdata$species,
# p = 0.7, # percentage of data as training
# list = FALSE)
#
# sub_testdata <- sub_trainingdata[-idx_train,]
# sub_trainingdata <- sub_trainingdata[idx_train,]
obj <- tune(svm, formula, data = sub_trainingdata, kernel="radial",
ranges = list(gamma = 2^(-8:2), cost = 2^(1:10)), probability = TRUE,
tunecontrol = tune.control(sampling = "fix", best.model = T))
classifiers_18c_16x[[i]] <- obj$best.model
cf <- caret::confusionMatrix(predict(obj$best.model, newdata=sub_testdata %>% select(-species)),
sub_testdata$species)
classifiers_18c_16x_plots[[i]] <- classifier_plot_svm(table = cf$table,
combination.nr = names(compositions_list)[[i]])
if(length(unique(sub_testdata$species))==2){
cf2 <- caret::confusionMatrix(predict(obj$best.model, newdata=sub_testdata %>% select(-species)),
sub_testdata$species, positive=as.character(unique(sub_testdata$species)[2]))
cf3 <- rbind(cf$byClass,cf2$byClass)
rownames(cf3) <- unique(sub_testdata$species)
cf$byClass <- cf3
}
classifiers_18c_16x_assess_plots[[i]] <- classifier_assessment_plot(cf, names(compositions_list)[[i]])
}
names(classifiers_18c_16x_plots) <- names(classifiers_18c_16x) <-
names(classifiers_18c_16x_assess_plots) <- comp_name6.3 Assessing of classifiers
There are mainly 4 measures:
Sensitivity: the probability that an individuals is classified as species X given that the it is of species X: P(Classified as species X | is of species X)
Specificity: the probability that an individuals is not classified as species X given that it is not of species X: P(Not classified as species X | is not of species X)
Positive Predictive Value PPV: the probability that an individual is of species X given that it has been classified as species X: P(is of species X | is classified as species X)
Negative Predictive Value NPV: the probability that an individuals is not of species X given that it has not been classified as species X: P(is not of species X | has not been classified as species X)
PPV is the most important for us!
library(patchwork)
for(i in 1:length(classifiers_18c_16x_plots)){
print(classifiers_18c_16x_plots[[i]] + classifiers_18c_16x_assess_plots[[i]] +
plot_layout(widths = c(4,2)))
}
7 Save classifiers
saveRDS(classifiers_18c_16x, "svm_video_classifiers_18c_16x_20220316_MergedData.rds")
# cl <- readRDS("classifiers_18c_16x.rds")